The Role of Stable Anatomical Landmarks in Automated 3D Model Superimposition: A Closer Look

Objective: To evaluate the concordance of automated 3D superimposition methods applied to digital models, with a focus on methods that consider stable palatal regions as geometric reference landmarks versus those that do not. Design and setting: This was a prospective, cross-sectional study using digital model files of patients undergoing orthodontic treatment in a university clinical setting. Participants: Sixty-one patients were prospectively enrolled and divided into three groups based on the type of orthodontic treatment they received: (20) non-extractive orthodontic treatment without intermaxillary elastics, (21) intermaxillary elastics, and (20) control subjects with no orthodontic movement. The inclusion criteria included the availability of complete pre- and post-treatment digital casts and the absence of significant craniofacial anomalies. Methods: Three superimposition methods were tested: (1) superimposition according to palate and palatal ridges, (2) best-fit superimposition of arches in occlusion, and (3) best-fit superimposition of individual arches. Discrepancies were identified by comparing the spatial positions derived from each method. Within three spatial axes, deviations of +/- 0.5 mm and +/- 1.15 degrees were not considered significant. Bland-Altman plots were used to quantify palatal rugae based and non-based spatial differences between methods. Differences in the superimposition results between the three patient groups were evaluated using ANOVA tests. Results: Differences in spatial position between the superimposition methods often exceeded the acceptable range. The results were compared between the three patient groups with a statistical significance of alpha = 0.05. In the present study, the high reliability of the superimposition method based on the palate and palatal ridges was observed. Conclusion: Superimposition methods based on the palate and palatal rugae provide superior accuracy in determining treatment-related changes in upper arch digital models. These findings illustrate the need for appropriate selection of superimposition techniques based on the study objective of using clinically relevant techniques.